Marburg Heart Score (MHS)

Clinical background

Chest pain is one of the commonest reasons for consulting primary care, and in most patients the cause is not cardiac. At the same time, coronary artery disease must not be missed, and the line between forgoing further investigation and referring for cardiac assessment is hard to draw on clinical estimation alone. The Marburg Heart Score (MHS) was developed specifically for the primary care setting, in which the prevalence of coronary artery disease is lower than in the emergency department and in which existing risk scores (developed for emergency care) cannot be transferred directly. The main purpose of the tool is to identify patients in whom coronary artery disease can be excluded with sufficient confidence, so that unnecessary referral and investigation can be avoided.

Calculating the Marburg Heart Score

The MHS consists of five variables, each worth 1 point:

$$\text{MHS} = \text{age/sex} + \text{vascular disease} + \text{exertional pain} + \text{not reproducible on palpation} + \text{patient's own belief}$$

The variables are defined as follows:

• Age/sex: women $\geq 65$ years or men $\geq 55$ years score 1 point
• Known clinical vascular disease (coronary, cerebrovascular or peripheral arterial disease): 1 point
• The pain worsens on exertion: 1 point
• The pain is not reproducible on palpation: 1 point
• The patient believes the pain is of cardiac origin: 1 point

The score ranges from 0 to 5. The threshold for not being able to exclude coronary artery disease is $\geq 3$ points.

The derivation study was conducted in 2005 to 2006 in 74 German primary care practices [1]. Consecutive patients presenting with chest pain were recruited (derivation cohort, $n = 1249$). An independent expert panel judged the presence of coronary artery disease from follow-up data at six weeks and six months. The variables were selected on the basis of adjusted odds ratios, and the scoring is equal (1 point per variable) because the predictive weight of the variables was roughly equivalent. Validation was carried out in a separate prospective primary care cohort. The AUC was 0.87 (95% CI 0.83 to 0.91) in the derivation cohort and 0.90 (95% CI 0.87 to 0.93) in the validation cohort [1].

Interpretation in practice

The MHS is primarily a rule-out instrument. The score determines whether coronary artery disease can be considered unlikely enough to forgo further cardiac investigation, or whether the patient should be taken further.

An important addition: the score complements but does not replace clinical assessment and the ECG. A normal score must not override a clinical suspicion of acute coronary syndrome, and ECG evidence of ischaemia should always drive management irrespective of the score.

Validation and performance

The first external validation was carried out in a German primary care cohort of 844 patients recruited from 56 general practitioners [2]. The AUC was 0.84 (95% CI 0.80 to 0.88). At the threshold of $\geq 3$, sensitivity was 89.1% (95% CI 81.1 to 94.0) and specificity 63.5% (95% CI 60.0 to 66.9). The negative predictive value was 97.9% (95% CI 96.2 to 98.9), confirming that the score is strongest as a rule-out tool. The positive predictive value was, by contrast, low at 23.3%, meaning that a high score does not confirm coronary artery disease with any certainty.

A comparative study examined whether the MHS adds anything to the general practitioner's immediate clinical assessment [3]. In a cohort of 832 patients, the general practitioner's unaided clinical judgement had a sensitivity of 82.9% and a specificity of 61.0%. The MHS had higher sensitivity (a difference of 8.5 percentage points) and similar specificity. When the MHS was used as a first-step triage tool, specificity increased by 11.6 percentage points at unchanged sensitivity. Best of all was the combination of the general practitioner's clinical judgement supported by the MHS, in which both sensitivity and specificity were higher than for unaided judgement. The conclusion was that the MHS as an initial triage tool can improve the clinical diagnosis of coronary artery disease in primary care.

A Dutch retrospective study from 2021 evaluated the MHS in a different setting: out-of-hours primary care [4]. In a cohort of 664 patients with chest pain, the c-statistic for the MHS fell to 0.77 (95% CI 0.69 to 0.84), compared with 0.84 to 0.90 in earlier primary care cohorts. At a threshold of $\geq 2$ (the study found that the optimal threshold in this population was lower than $\geq 3$), sensitivity was 81.3% and specificity 67.1%. None of the risk scores tested, including the MHS, performed better than the general practitioner's unaided judgement in this more acute setting. Only the INTERCHEST score, a newer rule based on pooled patient data from five countries, tended to perform somewhat better [4,5].

Limitations

The MHS was developed and validated for routine primary care, not for emergency departments or out-of-hours services. In the Dutch out-of-hours study, discrimination fell appreciably and the tool did not perform better than clinical estimation [4]. This is logical: the population prevalence of acute coronary syndrome is higher in the out-of-hours population, the symptom pattern is more heterogeneous, and the MHS variables do not cover acute features such as radiating pain, nausea or sweating.

The score is not validated for patients under 35 years, and the derivation cohort consisted exclusively of German primary care patients. External validation has so far been carried out in German-speaking primary care [2] and in the Dutch out-of-hours study [4], but is lacking in broader international primary care cohorts outside Europe.

A methodological weakness in several of the studies is the reference standard: the diagnosis of coronary artery disease was made by an expert panel from clinical follow-up, not systematically by coronary angiography in all patients. This may have led to misclassification, particularly for patients with mild or atypical coronary artery disease.

The positive predictive value is low even at a high score. A patient scoring 4 or 5 does not necessarily have coronary artery disease, and the MHS should not be used to confirm the diagnosis, only to determine whether it can be excluded.

References

1. Bösner S, Haasenritter J, Becker A, et al. Ruling out coronary artery disease in primary care: development and validation of a simple prediction rule. CMAJ 2010;182(12):1295-1300. PMID: 20603345
2. Haasenritter J, Bösner S, Vaucher P, et al. Ruling out coronary heart disease in primary care: external validation of a clinical prediction rule. Br J Gen Pract 2012;62(599):e415-e421. PMID: 22687234
3. Haasenritter J, Donner-Banzhoff N, Bösner S. Chest pain for coronary heart disease in general practice: clinical judgement and a clinical decision rule. Br J Gen Pract 2015;65(640):e748-e753. PMID: 26500322
4. Kleton M, Manten A, Smits I, et al. Performance of risk scores for coronary artery disease: a retrospective cohort study of patients with chest pain in urgent primary care. BMJ Open 2021;11(12):e045387. PMID: 34880006
5. Aerts M, Minalu G, Bösner S, et al. Pooled individual patient data from five countries were used to derive a clinical prediction rule for coronary artery disease in primary care. J Clin Epidemiol 2017;81:120-128. PMID: 27773828